Chief Executive Officer of ARM & Director, SoftBank
【完整公開】（原音呈現）COMPUTEX Arm CEO Rene Haas主題演講 重量級合作夥伴站台 @投資看非凡
#黃仁勳 #輝達 #Arm #AI #COMPUTEX
COMPUTEX2026今年主題演講由Arm CEO Rene Haas擔任，現場除了聚焦在代理式AI時代的關鍵基石，現場還有重量級合作夥伴嘉賓站台。
※下載非凡新聞ＡＰＰ看更多：https://reurl.cc/jr7A5p
※下載非凡商業ＡＰＰ學理財：https://reurl.cc/oYLj9j
更多內容 敬請鎖定 【非凡新聞台】【非凡商業台】
非凡電視台YouTube頻道
/ ustv
非凡電視台IG
/ ustvnews
非凡電視台粉絲團
/ ustv.fans
非凡電視台官網
http://www.ustv.com.tw
About Rene Haas
Rene Haas, CEO of Arm Holdings and a director at SoftBank, has been active in public discussions about the company’s role in the expanding AI infrastructure market. In May 2025, Haas commented on Intel’s decline, stating that the company was “punished” for missing mobile and for not investing in extreme ultraviolet (EUV) lithography at the same rate as TSMC, which he said made it “very difficult to catch up.” He also expressed optimism about U.S.-China collaboration on AI, saying that based on conversations he has had, China’s “minds are in the right space” regarding safety and guardrails, and that countries with capabilities need to “sit at the table to have the conversations.” In mid-2026, Haas discussed SoftBank’s plans to develop a large-scale AI data center, describing the company as “uniquely positioned to win” due to its assets in AI, robotics, energy, and compute. He argued that the demand for compute is not a sign of a glut, noting that “everyone needs more and more compute.” At COMPUTEX 2026 in Taipei, Haas highlighted the growth of agentic AI, stating that Arm believes “four times the number of CPU cores” will be needed in the same power envelope going forward. He also cited Google’s decision to move the head node of its TPU systems from x86 to Arm-based Axion chips, claiming a “60% less power at the same performance” benefit. Haas emphasized Arm’s reliance on Taiwan’s ecosystem, saying, “Without Taiwan there really is no ARM.”
Transcript (76 segments)
N
Narrator16:41
There's a specific kind of silence right before the world changes. It's that split second of realization. The moment when you know that life as you've lived it is about to become something entirely new. For more than 35 years, we've learned to recognize that moment. From Cambridge to the US to Taiwan to the world, we partnered with our ecosystem, solving problems together. We've been here from the start of AI, preparing the world for what's next. Powering intelligence everywhere in how we live, work, play, and move together. This is our moment. Because when you know you have the power to change everything, you step forward.
H
Host18:27
Please welcome ARM's Chief Executive Officer, Rene Haas.
R
Rene Haas18:41
Ni hao, welcome. I apologize for the delay, but we will get moving as quickly as possible. We have a lot of things to share with you this afternoon. June means Computex and June means a muggy evening and afternoon in Taipei. But it is wonderful to be back here. I think my first Computex was 2004, 2005-ish. So it's 20-plus years plus or minus when COVID hit. ARM started in 1990 and it was not long after 1990 that Taiwan and ARM started a relationship. Taiwan has built ARM. We are nowhere without the ecosystem and partners that exist inside Taiwan.
Now going back in time, we think probably around 1993-ish, a few years after we started, the first ARM chip was designed here. Those were early days. SoCs were a kind of a foreign thing. Design tools, physical design, EDA that could support SoC really didn't exist. But we were working with ETRI back in the day, who did some initial work with us to test out our IP, our methodologies. Not long after that, the first ARM chip manufactured in Taiwan. We believe it was TSMC. We're looking back. It may have been UMC. It was some early test chips. We didn't get into production really until later in the decade. But the first ARM chip was packaged and tested here. So really not long after ARM started, we were linked to Taiwan.
Some of the significant volumes though that really embody what ARM is all about started in the 2000s and this is before the iPod. If folks remember these little tiny MP3 players that had maybe 256 songs fit in your pocket from Creative, Diamond, Rio, companies like that. Those were all ARM-based. And of course then the iPod which in many ways was the catapult for the ARM technology being everywhere was here and that was in a chip designed by PortalPlayer that went into the very first MP3 player that took volume, which was the iPod. But it was really in 2008 when we had the revolution that was grown relative to mobile. The mobile revolution was really launched in Taiwan. Now, we were involved obviously with the early GSM phones as folks know from Nokia and LG etc. But it was really the launch of the iPhone and then the Android phones that soon followed and that revolution really launched the growth of ARM into a set of volumes we've not seen before. So it was really that period that was the most significant for us.
And today, and I'll talk more about ARM server CPUs, 100% of those CPUs are built here. And when we look in aggregate across everything that we've done in our history with all our partners, about 250 billion chips have been built in Taiwan, more than any other region in the planet.
I cannot tell you the gratitude we have as a company for the ecosystem, the people, the talent, the partners here. ARM is nowhere without the partners of Taiwan. Now, some very cool products have come out of the Taiwan ecosystem. When we look at the edge, products such as the Amazon Echo, OPPO, Vivo phones, Apple MacBooks, a product I use constantly, I don't mean this as a promo, but these Meta Ray-Ban glasses, they are amazing. I use them for phone calls, videos, messages, all here in Taiwan. Physical AI, the humanoids, the most advanced in the world, whether it's Tesla, Figure, Techmen, all the chips here built in the Taiwan ecosystem. And then of course cloud AI, whether it's the TPU racks, the racks by Nvidia, Graviton, everything here, as I mentioned, 100% of our ecosystem is built in Taiwan. So without Taiwan, there really is no ARM. Thank you again.
Now, what seems like a long time ago and in the world that we're living in with AI, we're living in light-year speed, we did an event called ARM Everywhere back on March 24th and at that time we were looking at what was going on relative to the growth of agents and Agentic AI. And at that time, and this is March 24th, not so long ago, showed a slide about the growth of OpenClaw relative to Linux and Kubernetes. GitHub stars on the left are exactly what you think they are. They are stars that rate the popularity or the stickiness of a certain application. OpenClaw reached levels almost beyond parabolic in terms of the takeoff. And this is back in March 24th. And what that told us was that the growth of these agentic platforms were driving demand for CPUs in a way we had not seen before.
And the logic behind that is quite simple. GPUs, XPUs are amazing at generating tokens. That is their purpose. Whether it's training to generate the learning or inference to deliver the tokens, the token machine, the token factory is the accelerator. But agents, unlike humans, don't sleep. And agents beget agents that beget agents. And all of those tokens that need to be distributed, managed, orchestrated, delivered to the destination, that's only a workload that CPUs can do. CPUs of course in conjunction with a full system design. So we made a comment back on March 24th and I think we were probably one of the very first to do this that said we believe going forward that four times the number of CPU cores needed in the same power envelope going forward. Now that multiplier I end up getting so many questions relative to show me the math and how do you figure that out and not long after that you started hearing numbers of 4x, 8x, 10x. It's a hard number to predict just based upon the growth rate of these agents but what we do know is as follows.
If we look at today what we're seeing in terms of agentic growth, even fast-forwarding from the 24th of March, this is just exploding. We're seeing this with SaaS companies, whether it's Snowflake or Salesforce or ServiceNow, who are developing all the agents relative to running in the back lane. The explosion of Anthropic with Claude Code, Codex from OpenAI, all these agentic workloads are driving in more demand. And what that does in turn is drive a very, very significant growth. Clicker is doing here in terms of where the CPUs go. So if you change the axis on the Y side to units and you look forward in terms of what the growth rate looks like, CPUs are even growing faster than we had thought. And we are seeing this across the board. It's not just ARM. Of course, I'll be promoting ARM a little bit more later, but we're seeing this from everyone who's in the CPU business. The demand for these CPUs continues to explode because the agents beget agents beget agents.
Now, is the number 4x? Is the number 6x? Is the number 8x? I don't know. But what I do know is that it's getting bigger. That the agents continue to accelerate relative to the growth and with that CPU growth is also raising. We threw out a number back on March 24th around a CPU TAM in 5 years going to north of $100, $120 billion. And again, at the time when we did that event, we had a lot of questions from media, investors, analysts saying that number seems a little too aggressive. Not sure how you got there. Fast forward, the numbers that people are talking about are almost twice that number, if not larger. What we do know is that AI, agentic workloads, because of the more tokens you generate, the more information that's being used, the more that they are agentic, drives demand for compute. And of course, we have an answer for that. The ARM Neoverse CPU.
Now, this CPU, as I mentioned before, 100% built in Taiwan. I'm going to show you a video that we showed on March 24th. Going to show it to you again for those that didn't see it. I want to show it again because frankly, I love it. It's a great video and says just about everything you want to know about the product, but also emphasizes the importance of the Taiwan ecosystem.
I think I can watch that video every single day. I just get so motivated, enthused by what I see there. So the ARM Neoverse CPU, built in Taiwan. TSMC, our partner, we are now in production of this product. One of the things that we emphasized early on when we talked about potentially delivering solutions into the marketplace was that we didn't want to talk about the product until we had customers, the product was shipping and equally as importantly that we had partners who could help deliver the product to market. We understand that in this world it's not just about delivering a chip, but it's delivering a full system with partners. And we've worked with some of the best on the planet all here in Taiwan. I understand there are actually some that are out there in the demo area. I think we may even have a full rack I've heard from Super Micro sitting out there. But whether it's ASRock or TSMC, our FAB partner, Quanta, Ingrasys, Super Micro, Aspeed, all fantastic partners who enabled our ecosystem to deliver amazing solutions.
Now, this product comes in two flavors from a system standpoint and one of the things that we really emphasize with the ARM Neoverse CPU is maximum performance, density, and efficiency. Of course, our hallmark is around energy efficiency. We were born from mobile phones. We designed a custom CPU way back in the day that had to fit into a plastic package and run off batteries. And that is a mindset that sits inside our engineers in everything that we do. And it translates to amazing solutions and products. An air-cooled rack, 36 kW, 8,000 cores, and a liquid-cooled rack that has over 45,000 cores, 200 kW. So, two different solutions. But what's key about this product line is the performance per rack, performance per watt. Two times the performance per rack versus a comparable x86 system. Basically means same power envelope, two times the benefit in terms of performance. If you want half the power, you still have equivalent performance. So it's incredibly efficient.
But more importantly, when you think about what goes into these giant data centers and we're seeing announcements literally daily. In fact, the parent company of ARM, SoftBank, just announced a partnership in France for a 5-gigawatt data center. These data centers are incredibly capital intensive. The energy costs are huge. So having the benefit of performance per rack, more CPU in the same power envelope has huge, huge benefits versus the competition. We estimate about 10 gigawatts of capacity, over 10 billion, up to 10 billion of savings. But as we go forward and we have more and more CPUs inside the systems, you'll get even more benefit relative to using the ARM Neoverse CPU.
Now, we were super proud back in March to talk about our partners, people who had embraced the solution, customers that we had signed up, Meta, Rebellions, SAP, Cerebras, OpenAI, SK Telecom. And that was just on March 24th, and we talked about our customer base and who had adopted the product. I'm proud to say that since that time, even more companies have joined the family. Oracle, huge partner with OCI. We have a long history with Oracle. They've now joined the Neoverse CPU family as well as ByteDance. Two new partners part of the family validating that the Neoverse CPU solves real world problems.
Now, we talked about this back in March and I want to emphasize it again. We are now a full end-to-end solution provider. So while we do have production silicon of the ARM Neoverse CPU, not everyone wants to buy the ARM Neoverse CPU and that's okay. We have compute subsystems, many partners who take that, and we have just standalone IP in this space whether it's Google, whether it's Amazon, whether it's Nvidia, whether it's Microsoft, we have many, many customers who are on the left-hand side of that slide and we intend to provide solutions to whatever the customers want to see. And that is very important because the momentum is really increasing for us now with Agentic AI and whether it's our own CPU or our partners. Very significant announcement took place last month where Google announced for their TPU 6e and 6i that the head node, the CPU that interfaces into the accelerators, is going to move from x86 to Axion, which is their internal chip using ARM Neoverse. 60% less power at the same performance.
Andy Jassy had a great quote I think one of the earnings calls that basically said for Graviton, we had two customers say can we buy everything that you have? Graviton now has more than half of their design starts are based on Graviton versus x86. From a few years ago that was zero. And of course, Nvidia who announced Vera, amazing partners. Vera is an amazing product. The list of partners is far larger than here. I didn't have a slide big enough for all of them, but Nvidia's had a tremendous momentum with Vera.
Now, our intentions are very clear for the ARM Neoverse CPU. We intend to be in this for the long term. It's multi-generational. ARM Neoverse CPU 2 is already underway. And as you can imagine, it has more cores, more power efficient, better performance. And ARM Neoverse CPU 3 is on the way. But these are all based on the compute subsystems that we intend to deliver along with the chips and they'll be lined up roughly on the same cadence. So the CSS's that we deliver to our partners, those are what we use to enable our end devices. So that's ARM Neoverse CPU, which has had incredible momentum.
Now I want to switch gears a little bit because Computex to me always, having come here 20 years ago for the very first one, was always about the old exhibition hall, floppy disk controllers, USB cables, all kinds of things in terms of IT malls. And you could go into these shops and buy almost anything under the sun. It was like a mini Fry's, 20 of them on a floor in a building that was 10 stories high. And that's how people bought PCs back in the day in terms of how they shop for them. And if you think about how these PCs were built and how we used to buy, it was very interesting. You'd have literally every single price point you could think about, whether it was a base entry laptop, raise your hand if you remember the netbook. I knew the Nvidia guys would remember that one. We have battle scars from that one. All the way up to high-end gaming machines. But literally, these units were priced at $50 price points. You had feeds and speeds, clock frequency, memory size, etc., etc., and everybody was trying to position for the slice of the pie.
So much has changed obviously, not only in how we buy PCs, but more importantly, how we use these products. How we use the products has really, really evolved with obviously what the smartphone has done, what the web has done, what applications have done. And what we see is that they've really started to bifurcate into kind of two areas I would say. One is, and I think many of you can identify this on the bottom left, is I need a machine that is on the go, battery life is really good, connects everywhere, and I need it to kind of look like a large phone with a keyboard where I can do work, but it maps very closely to what my phone does. And if I think about myself personally, I have one of these flip phones which I use for reading documents and reviewing presentations. And I'm a CEO so I create very little these days. I review many things. But what I find is I go back and forth a lot between that smartphone that flips like a tablet into the PC. But it's really super important that the PC and phone are synchronized and they can do things back and forth very, very quickly.
There's also an extreme performance workload and that is I'm either running agents, I'm either running models, I'm doing some development work, I need some very, very extreme level of performance. So there's really two different components in two different areas in terms of how they all work. So only ARM really enables this for PCs and I think that's a very, very key distinction in terms of the way we used to think about this category back in the day where literally you had every single price point covered, every single feed and speed. Now you want two different ends of the spectrum and whether it's long battery life, great AI experience, we're in that bottom category. But if you also want the agentic type of performance, we're there as well.
Now, specifically when we look at the units that are there, you can see that you've got the Acer device, Mac Neo, pretty interesting product, the Chromebook, Microsoft Surface, Mac Studio, of course, the Nvidia RTX Spark, which was just announced, which I'll talk about. But these two broad categories are very unique to ARM. And I get lots of questions, you know, over the years about Windows on ARM and when is ARM going to really take place to be a significant player in laptops and the compute space. I would argue that we are now actually there because when we look across the spectrum of the operating systems that are supported whether it's Linux, whether it's macOS which is 100% on ARM today, Chrome, Windows, only ARM can enable this across the board and this would not be done without huge, huge, huge cooperation from all of our partners who are listed there, the folks on the operating system side that we work so closely with. We've worked for decades with Apple. We've worked for decades with Google and Microsoft. This work does not happen overnight. There is a huge amount of effort to go off and make this happen. And I want to give an applause and thanks to all of our partners to make this work.
Now I want to talk about a product that we knew was being worked on and we are proud to be a partner with Nvidia on the RTX Spark powered by ARM. 20 cores, ARM-based cores in the custom Grace CPU. I believe that is the most CPU cores that you can find in a laptop anywhere. But when you pair it with Blackwell, the world's most powerful GPU for AI, you have an incredibly special product. One petaflop of FP4, huge amount of memory, full Windows native on ARM. Amazing product.
And of course, as you'd suspect, partners who were there already, Acer, Asus, Dell, Gigabyte, HP, Lenovo, Microsoft, MSI, I think I saw a Surface Ultra that was announced. An amazing product. Congratulations again to the Nvidia team for making all this happen. Now, our role here was working very closely with Nvidia and with MediaTek using our CSS strategy. And again for those who are not familiar with what our compute subsystems do, the CSS is basically the building blocks that we use to put together everything to build a full end solution system. The CPUs, the GPUs, the system IP, the memory controllers, everything that goes into building a custom SoC. We provide these to our customers. We did this with MediaTek as either full solutions they can take or building blocks that they can start with. So we see a very significant opportunity again given the strategy we talked about with IP and compute subsystems around the ARM Neoverse CPU, very, very similar with what we're doing with the CPUs for the CSS's and I think the PC space is going to be a very, very interesting domain as I said going forward because with these use cases on the bottom left again the kind of use that I am relative to using the systems for creation and things of that nature. The high-end systems when we start thinking about where agents can go and how agents interface with us, it's going to be a very, very different domain and I think this product from Nvidia has really demonstrated its capability.
So, I'm not sure if the systems are available yet, but we actually got access to some of the hardware and technology and we decided to give it a spin. Complete surgeon's general warning here. The following video was AI generated. So, please don't have your legal teams contact us. But let's take a quick look.
Now, I know you're probably saying, 'I'm not sure that's AI because the dude always wears the same clothes.' But on the other hand, those are events that I would not actually do myself, but I think it's just a small example of the kind of creation that can be done, you know, on these computers and where I think we're going to go with Agentic AI. Now, I want to be able to talk more about the product, but I'm kind of thinking that there's probably someone better than me to join me on stage to talk about the RTX Spark and everything that Nvidia does. So, I'm going to introduce a special guest here. If my clicker behaves.
J
Jensen Huang48:45
That's a pretty cool video of Rene. Superstar.
R
Rene Haas48:51
Action hero.
J
Jensen Huang48:52
Do you buy... not just a superstar, he's an action hero?
R
Rene Haas48:55
I think the night market was the part I thought was the coolest.
J
Jensen Huang48:58
Yep. Well, that's the most exciting part of your video.
R
Rene Haas49:01
Yeah. Well, thank you for joining. I appreciate it.
J
Jensen Huang49:05
Yeah. So,
R
Rene Haas49:06
So tell me Jensen, congratulations on the RTX Spark. Amazing.
J
Jensen Huang49:12
Thank you.
R
Rene Haas49:12
Windows on ARM is not a new thing. Why is this one going to be different?
J
Jensen Huang49:17
Look at his stock price. I announce a product. Look at his stock price. Every product I announce, his stock price goes up. Nothing happens to mine.
R
Rene Haas49:35
Let's also stay for the record that... I'm very happy about that. Let's also stay for the record that you were a shareholder and you sold.
J
Jensen Huang49:43
Yeah. Yeah. Well, I needed the cash.
R
Rene Haas49:49
So, how... what were we talking about?
J
Jensen Huang49:51
RTX Spark.
R
Rene Haas49:52
RTX Spark. How is it going to be different this time?
J
Jensen Huang49:55
Well, we wanted to reinvent the computer. You know the PC has been here for 40 years and the operating system code written by hand is now going to be replaced with applications that are agentic. Now these agentic systems, agentic AIs will use the PC, will use the tools in the PC. And so when we imagined this future we thought let's see how would we change the architecture and how would we change the operating system and reinvent the computer and that's kind of where we are. And so one of the things that we realized is that an agentic system really wants to have excellent CPUs which is the reason why we used ARM and it has a 20-core CPU, has to have excellent single-threaded performance, the parameters, the memory has to hold a lot of parameters and so we created a new numerical format called NVFP4 so that we can compress the large language models as much as possible and fit a very smart AI into the system memory. We also wanted to unite CUDA that is for accelerated computing and CUDA cores, tensor core processing into one processor. And the reason for that is because when you're operating these agents and they're thinking and they're using the tools, the agents are fast. And when the agents are fast, they expect the tools to respond quickly. And so that's why we're accelerating all of the tools. We're accelerating Adobe. Adobe announced they're going to rearchitect Adobe Photoshop and Premiere so that it's CUDA accelerated and agentic-ally accessible and so we're accelerating applications. We accelerated Blender with RTX. We accelerated, you know, we're going to accelerate everything. We accelerate Adobe, Autodesk, Siemens, we're going to accelerate every tool and once these tools are accelerated then they can respond to the agents very quickly. And so we now in order to build this computer, this SoC, unless you have the ability to integrate with the CPU and adapt the CPU to exactly the shape of the computer it's really quite impossible which is the reason why ARM is perfect.
R
Rene Haas52:08
Well thank you and when you think about the agents running locally...
J
Jensen Huang52:12
And the key word there is ARM is perfect.
R
Rene Haas52:20
The other key word is thank you. Agents running locally versus agents running in the cloud. How do you think about that as a trade-off and where do you think that goes over time?
J
Jensen Huang52:44
Well, you know, when ultimately these computers, these personal computers are going to be becoming agents that are running all the time. They're autonomously used, running all the time. I could imagine I today if I left my laptop at home where I left my laptop in the hotel, I won't use it again until I get there. But in the future, you just pick up your phone and you chat with your agent. You're chatting with your PC in the future and that you maybe there's something that you needed to have done and sent to you. Maybe there's a speech I need to have quickly written. And so, you know, I'll be working with my agent, working with my assistant, and that is now the ARM personal computer, right? And so the PC is working in the back while you're not there. It's working. Yeah. And so, if I want to do something that requires a cloud API, of course, I'll call it into the cloud API. But whatever locally, we're going to continue to do on the PC, which is kind of the nature of PC. The nature of a personal computing device is that whatever you can do on the device, you do. You don't have to worry about metering. You don't have to worry about the time spent, but whatever you need to do in the cloud, you will.
R
Rene Haas53:53
And when you think about the complexities of the models, do you think PC performance and architecture can scale? I mean, you guys are doing incredible work with where Blackwell and then Rubin, etc. How do you think that all maps together in terms of scaling the systems?
J
Jensen Huang54:05
Well, if you look at the RTX Spark PC, it's got 128 GB memory. If it was completely compressed into NVFP4, then you can have a 100 billion parameter model working on your PC all the time. And a 100 billion parameter open model, say Nemotron 3 Super, say that's a really, really good model. And so it could do a lot of the basic work and whatever deep thinking and frontier model that you need to use, it's just connected to cloud anyhow.
R
Rene Haas54:36
Do you think that changes what happens in the cloud in terms of this classic client-cloud model? Do I need as much compute in the cloud versus on the client or do you think there's just so much compute that needs to get done?
J
Jensen Huang54:49
They just... these agents are going to be, you're going to be, you have agents and sub-agents and teams of agents. They're going to be working in the cloud. They're going to be working on devices. And so it's just like today in a lot of ways. Mobile-cloud is not cloud only, not mobile only. It's mobile and cloud. And so it allows you to have a really great personal computing experience, you know, your own experience. But whatever you need to connect to the cloud, you will.
R
Rene Haas55:16
And so do you think, and this may be a bit of a provocative question, but as these agents are running in the background and they're doing a lot of the work, does the operating system matter? Is the agent really the OS, if you will, and it does the work and isn't so reliant on the OS? Where do you think that goes over time?
J
Jensen Huang55:32
Well, the operating systems are going to be just as important as ever before, if not more important. And the reason for that, and this is the controversial part that people say AI comes along, software is dead. Nothing is further from the truth. And now people are starting to realize that when agents are here, they're going to use tools. And so those tools are more important than ever. And so they're going to use Adobe Photoshop, they're going to use Adobe Premiere, they're going to use Canva, they're going to use the, you know, they're going to use the Siemens tools, they're going to use, you know, tools, whatever they have on the device. This is the incredible part today. Most of us probably know 10, 15, 20% of the features of a tool. If you know how to use Photoshop, you know, use Lightroom, unless you're an expert like my son, it's kind of hard for you to know all of the features. But now with your agent, you tell the agent what you're looking for and the agents know exactly how to use the tools because it's read a skills file. It's essentially read the manual of that tool. And so now it goes and uses the MCP or the CLI connected to that tool and it does everything you need to do. It's going to unlock all these tools. These tools are going to be more useful and more valuable than ever. And that these tools run on the operating system. So, we're going to need Windows. We're going to need, you know, all these APIs and all these tools for a long time.
R
Rene Haas56:57
So, Nvidia is involved, understatement, in everything around AI. I mean, you guys do everything around the networking, the systems, you know, where all the bottlenecks are. When you think about over the next number of years, where are the constraints to growth? Where do you think they are?
J
Jensen Huang57:16
Well, it's probably going to be everywhere. This is, at this point if you look at our evolution, first Hopper was designed for training. Then Grace Blackwell was of course great at training, but we also specialized NVLink 72 for inference. And at first people thought, you know, inference was easy. And we explained to people that large language models and to be able to inference very quickly and generate these tokens as efficiently as possible, you're going to need a very complicated computer. And so Grace Blackwell NVLink 72 is the most efficient and we produce the lowest cost tokens in the world. And so that was a big breakthrough and now people understand that you want very advanced systems to generate tokens at very low cost. Vera Rubin took, of course, all of that and we extended it to run agents. At first, when I said that two years ago, most people had a hard time understanding what that meant. But now they realize that an agent is orchestrating thinking, is using tools, it's accessing long-term memory, it's dealing with short-term memory, working memory, and it's doing memory compaction to remember to think about what should I remember for the future? How do I index SQL memory? How do I index structured memory? How do I index unstructured memory? And so how do I deal with all of that? That agentic system is what Vera Rubin is. And it's a large system. And so people are now starting to understand that when we were thinking about agentic systems, we were really thinking about a new computing application pattern and that it really requires a new architecture. Well, now the big breakthrough of course these agents now are producing useful AI and that's the reason why all of our growth, your growth, my growth is just so incredible because when AI becomes useful then the tokens that are being generated are profitable and when token generation is profitable everybody wants to generate a trillion times more tokens. The other part is that the agent, the application, this agent compute pattern is a thousand times, maybe a hundred thousand times and depending on the work it's a million times more than chatting. And so you could see that the agents are working, they're working for minutes, sometimes days, sometimes weeks and so instead of a chatbot which responds from one click now the AI is thinking, using tools, reading, thinking some more, planning, trying and so the amount of tokens that we have to generate has increased tremendously. The profitability of the tokens obviously is driving demand. So the compound effect of need more compute with more demand that compounded effect is what you and I are experiencing. And so we're seeing constraints almost everywhere. In our case, we were fortunate that we planned, one of the best things about ARM is that they don't have to worry about the supply chain. The supply chain of IP is electrons and you could use as many electrons as you need. And so I love his business model. I mean, as you know, I tried to buy it.
R
Rene Haas1:00:43
I tried to...
J
Jensen Huang1:00:45
I tried to become ARM, you know.
R
Rene Haas1:00:47
We were willing. I was trying to become Mark.
J
Jensen Huang1:00:51
Were you really? Rene and I used to work together and then we tried to work together again. But anyways, that was okay. I'm sad still. I'm a little sad. But this is a happy meeting. So my point is in our case, we saw agents coming and we saw Vera Rubin coming. So we did a good job planning our supply chain and so our supply chain can support our very robust growth. We grew almost 100% year-over-year this year. We're going to grow very aggressively next year. And so our supply chain could support our growth, but the fact of the matter is demand is even higher than that.
R
Rene Haas1:01:28
Yeah. I was talking with CC and Kevin this week and they were saying, you know, at some point gravity has to take over. They've never seen four consecutive years of a semiconductor cycle that looks this good. But when you look at the things that you just described, there's no reason it can't continue in terms of fundamentals.
J
Jensen Huang1:01:46
That's right. Take a step back. What's happening? Take a step back and think what's happening. What's happening is the computer industry was limited by the number of people using the computers. And now we have agents that are autonomously using computers. And so we're going to have instead of 1 billion humans using computers, we will have tens of billions, maybe more than that of agents and robots and self-driving cars using computers. And so the question is how large can the computer industry be? And so my sense is that at this point it's a foregone conclusion that what is a trillion dollar, multi-trillion dollar industry is likely 10 times larger. And so we're on our way to...
R
Rene Haas1:02:29
And that's why Nvidia is the largest market cap company in the world. And if you combine the two companies we'd be the largest in the world still.
J
Jensen Huang1:02:36
I love that. I love that. That's such a great idea.
R
Rene Haas1:02:42
So, you know, thank you. Congratulations on RTX Spark. Just amazing.
J
Jensen Huang1:02:46
Well, congratulations on everything you guys are doing.
R
Rene Haas1:02:48
I have a small gift for you.
J
Jensen Huang1:02:50
Really?
R
Rene Haas1:02:51
Someone's going to give here. So, for those who may not recognize what this is, and I'm going to sign it. This is very, very real, by the way. The very first... Jensen talks a lot about resiliency and sticking with things. Tegra 3 was the first Windows on ARM laptop that was announced.
J
Jensen Huang1:03:13
How come? How come when we were younger? I have to tell you, I think I aged better.
R
Rene Haas1:03:29
Do you guys agree? I think I feel I aged pretty well. Come here. Here. You're my guest. It's to you.
J
Jensen Huang1:03:39
I sign it back to you. It's got treasure.
R
Rene Haas1:03:42
No, you sign it back to me. There's a contract. There's invoices.
J
Jensen Huang1:03:47
We can't do that. We know that, Dave. All right. Thank you very much. Thanks, guys.
R
Rene Haas1:03:55
Buy ARM. Buy ARM. I try.
One of those things was real up there. That actually was a real system that we worked on and Fish and Cows, those guys will remember on that. I think I aged a little bit better than he did by the way. So to wrap up, one compute platform, cloud to edge, showed these products before, it's the ARM AI compute platform that enables systems from the very, very smallest to the very, very largest. And we do this through a very consistent effort with software. 22 million developers, the largest developer community across the planet for any compute platform. But as I said, none of this happens without incredible cooperation and dedication from our partners. And again, I just want to say thank you to Taiwan. ARM is nowhere without Taiwan, the ecosystem, the people, the engineers, the supply chain managers. Thank you so much for everything you've done. Thank you for attending today.